The near surface remote phenology is a new area of phenology and offers great opportunity to obtain an impartial assessment of seasonal changes in tropical environments. In this project we aim to incorporate new technologies of phenological observations using digital cameras installed at the field, promoting the analysis of the RGB color channels : red ( R) , green (G ) and blue ( B ) and test them as an innovative tool and supplementary to the direct on-the-ground phenological observations data of leaf flush and leaf senescence in tropical environments . This proposal is part of the project e-phenology, recently approved by FAPESP - Microsoft Research Program: "Combining new technologies to monitor phenology from leaves to ecosystems." The goals of this project are: (i) analyze the remote phenology of species and community area of a cerrado sensu stricto (Core area), correlating to climate data , and validating with data collected on-the-ground and identifying functional groups for different species based on their leaf strategies; (ii) monitor and describe phenological patterns for different vegetation sites along a gradient of seasonality using data from digital cameras and correlate them with local climatic data, to identify environmental triggers in different vegetation types; (iii) identify a color change pattern of the RGB color channels for leaf exchange at the community and species levels within environments that are under different seasonal pressures; (iv) refine the analysis of phenological responses of the cerrado species in the digital images through analysis of leaves spectral information in order to more accurately categorize the patterns of leaf exchanges at the cerrado vegetation and the importance of the red channel. The sites selected within a seasonal gradient includes: outcrops, cerrado (grassland), cerrado sensu strito, semi-deciduous forest and Atlantic forest. This time series obtained for each one of these vegetation types will be analyzed and correlated with local climatic data. Thus, we can describe phenological changes detected by digital cameras for the community, correlate these changes with abiotic factors and verify how this response may vary for different vegetation sites. This work will offer a new tool for phenological monitoring in tropical vegetation. (AU)
News published in Agência FAPESP Newsletter about the scholarship:
PEDRONETTE, DANIEL C. G.;
ALBERTON, BRUNA C.;
MORELLATO, LEONOR PATRICIA C.;
TORRES, RICARDO DA S.
Unsupervised Distance Learning for Plant Species Identification.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING,
n. 12, 1, SI,
Web of Science Citations: 4.